Sex-specific immunization for sexually transmitted infections such as human papillomavirus: insights from mathematical models

Abstract

Background: Sex-specific differences regarding the transmissibility and the course of infection are the rule rather than the exception in the epidemiology of sexually transmitted infections (STIs). Human papillomavirus (HPV) provides an example: disease outcomes differ between men and women, as does the potential for transmission to the opposite sex. HPV vaccination of preadolescent girls was recently introduced in many countries, and inclusion of boys in the vaccination programs is being discussed. Here, we address the question of whether vaccinating females only, males only, or both sexes is the most effective strategy to reduce the population prevalence of an STI like HPV.

Methods and findings: We use a range of two-sex transmission models with varying detail to identify general criteria for allocating a prophylactic vaccine between both sexes. The most effective reduction in the population prevalence of infection is always achieved by single-sex vaccination; vaccinating the sex with the highest prevaccine prevalence is the preferred strategy in most circumstances. Exceptions arise only when the higher prevaccine prevalence is due to a substantially lower rate of natural immunity, or when natural immunity is lifelong, and a prolonged duration of infectiousness coincides with increased transmissibility. Predictions from simple models were confirmed in simulations based on an elaborate HPV transmission model. Our analysis suggests that relatively inefficient genital transmission from males to females might render male vaccination more effective in reducing overall infection levels. However, most existing HPV vaccination programs have achieved sufficient coverage to continue with female-only vaccination.

Conclusions: Increasing vaccine uptake among preadolescent girls is more effective in reducing HPV infection than including boys in existing vaccination programs. As a rule, directing prophylactic immunization at the sex with the highest prevaccine prevalence results in the largest reduction of the population prevalence.

Figure 1. The differential impact of sex-specific immunization on the reproduction number and on the prevalence of a heterosexually transmitted infection.

(A) The effect of immunization coverage among females (v _f) and males (v _m) on the projected reproduction number in a partly vaccinated population R_v. (B) The effect on the equilibrium prevalence of infection among men, I _m, and women, I _f. Darker colors correspond to lower values; the region where R_v<1 corresponds to I _m+I _f = 0, i.e., elimination of infection from the heterosexual population. In this example, R ₀ = 3.45 and women have both a slower recovery from infection and a lower probability of transmitting infection than men. The largest reduction in the reproduction number is achieved by allocating all vaccine to a single sex; the choice between vaccinating males or females is arbitrary. The largest reduction in the equilibrium prevalence of infection is achieved by allocating all vaccine to females, for any given coverage below the threshold required for elimination.

Figure 2. The immunization coverage in a girls-only vaccination program below which vaccination of boys is more effective in reducing prevalence.

(A) The threshold coverage for combinations of recovery rate α_f and α_m given equal transmission probabilities β = 0.9. (B) The threshold coverage for combinations of transmission probability β_f and β_m given equal rates of recovery α = 0.1. Contact rate c = 1 and death rate d = 0.02 deaths per year. The set of parameters for which male vaccination is an attractive option becomes increasingly restricted with higher female immunization coverage. Vaccinating males is rarely attractive if at least 40% coverage has been achieved among females.

Figure 3. The effectiveness of HPV vaccination depends on which sex is being vaccinated and on the heterogeneity in sexual behavior.

(A) The equilibrium prevalence of HPV16 infection in relation to immunization coverage by vaccinating girls only, boys only, or both girls and boys at an equal rate. (B) The equilibrium prevalence of HPV16 infection in relation to female immunization coverage for various assumptions of heterogeneity in sexual behavior. Results in (A) assume three dynamic activity classes plus an age-specific partner preference function. The default parameters were obtained by fitting this model to prevaccine data on HPV16 infection in the Netherlands .

Figure 4. The effectiveness of male-only or female-only HPV vaccination depends on sex-specific differences in viral transmissibility and natural immunity.

(A) The equilibrium prevalence of HPV16 infection in relation to immunization coverage for different assumptions regarding viral transmissibility. (B) The equilibrium prevalence of HPV16 infection in relation to immunization coverage for different assumptions regarding natural immunity. Solid lines represent a strategy of vaccinating preadolescent girls, and dotted lines represent a strategy of vaccinating preadolescent boys. Natural immunity is lost over time at a rate κ (per year). Default parameters, β = 0.8 and κ = 0.04 for both sexes, were obtained by fitting this model to prevaccine data on HPV16 infection in the Netherlands .